A Hybrid Deep Learning Approach that Uses Spatiotemporal and Behavioral Data to Predict Students' Academic Performance

The exponential growth of educational data necessitates innovative approaches to mining and utilizing this information to enhance educational practices. This study proposes a hybrid deep learning framework for educational data mining (EDM) that integrates various data sources, advanced feature engineering techniques, and state-of-the-art classification algorithms to improve learning outcomes and institutional decision-making processes. The research utilizes a diverse dataset comprising EDM applications, real-time educational data, and synthetic student data to develop robust models. Feature engineering is conducted using a hybrid approach that combines TF-IDF, N-gram, bigram relational models, autoencoders, and density-based techniques, aiming to maximize data representation and reduce dimensionality. The classification phase incorporates an array of traditional and deep learning methods, including Naïve Bayes (NB), Support Vector Machine (SVM), Artificial Neural Networks (ANN), Random Forest (RF), AdaBoost, Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and a novel hybrid RNN-SVM model. The proposed hybrid RNN-SVM classifier demonstrates superior accuracy and scalability by leveraging RNN's sequential learning capabilities and SVM's margin-based classification efficiency. Additionally, a recommendation module is designed to provide actionable insights, including class improvement strategies and industry-oriented suggestions, thus bridging the gap between academic performance and professional readiness. The hybrid deep learning framework not only enhances predictive accuracy but also facilitates informed decision-making for educators and policymakers. Experimental results validate the framework's efficacy in mining meaningful patterns from complex educational datasets and optimizing learning strategies. This research highlights the transformative potential of hybrid deep learning in advancing the field of EDM and fostering improved educational practices.

Mental health disorders like depression and anxiety pose global challenges, requiring accurate, non-invasive detection methods. Classical modes of diagnosis are typically based on self-reported symptoms or clinical evaluation, which could be subjective and protracted in time. To address these limitations, this study proposes NeuroHAGWO-Net, an advanced artificial intelligence-based framework for automated mental health status detection using multimodal data. The proposed model integrates electroencephalogram (EEG) signals and behavioral textual data to enable early and reliable mental health screening. EEG signals are pre-processed with Empirical Mode Decomposition (EMD) for noise removal, while behavioral text data is transformed into embeddings using Bidirectional Encoder Representations from Transformers (BERT) models. The hybrid BiLSTM-CNN architecture captures temporal dependencies and spatial patterns in EEG data, enhanced by integrating behavioral embeddings for multimodal analysis. Features are selected using a novel Hybrid Ant-Grey Wolf Optimization (HAGWO) approach, combining Ant Colony Optimization (ACO) and Modified Grey Wolf Optimization (mGWO), respectively. The AI-based mental health detection is performed using NeuroVisionNet, integrating EfficientNetV2 and Temporal CNNs (T-CNNs). The model's performance is validated on two datasets: behavioral data and EEG signals data. On behavioral data, it achieves an accuracy of 0.9945, precision of 0.9874, sensitivity of 0.9935, specificity of 0.9915, F1-Score of 0.9909, Matthews Correlation Coefficient (MCC) of 0.9925, Negative Predictive Value (NPV) of 0.9905, False Positive Rate (FPR) of 0.0151, and False Negative Rate (FNR) of 0.0092. With its strong accuracy and efficiency in detecting mental health situations under diverse data modalities, NeuroHAGWO-Net Model proves to be a robust tool for early mental health screening and clinical support using modern optimization techniques and deep learning architectures.

Accurately predicting student performance and identifying anomalies in academic datasets has become increasingly crucial for enhancing educational outcomes and enabling data-driven interventions in modern learning environments. Traditional statistical methods and conventional machine learning approaches often struggle with the multidimensional nature and increasing scale of contemporary student datasets, which encompass diverse academic, behavioral, and socio-demographic variables. This study explores advanced deep learning techniques; including Autoencoders for unsupervised anomaly detection, Recurrent Neural Networks with Long Short-Term Memory architectures for temporal pattern recognition, and Deep Neural Networks for comprehensive performance prediction to address these challenges. The proposed framework demonstrates significant improvements in detecting subtle performance anomalies that often precede academic difficulties, while simultaneously predicting longitudinal success patterns with greater accuracy than traditional methods. By leveraging the hierarchical feature learning capabilities of deep architectures, our system enables early identification of at risk students through continuous analysis of complex, nonlinear relationships in educational data, allowing institutions to implement timely, personalized interventions. Research studies have empirically validated the effectiveness of these models in educational contexts, showing superior performance in measuring student achievement patterns and predicting learning outcomes. The findings contribute to theoretical advancements in educational analytics but also provide practical insights for curriculum designers and policy makers seeking to optimize instructional strategies. Furthermore, the study establishes significant benchmarks for educational contexts by demonstrating how deep learning can enhance both teaching methodologies and student support systems through data-driven insights. This research makes a substantial contribution to the growing field of Educational Data Mining by proposing a robust deep learning framework that serves as both a predictive tool and a baseline for future studies in student performance analysis, while also addressing critical challenges in model interpretability and implementation scalability within real-world educational settings. References Abatal, A., Korchi, A., Mzili, M., Mzili, T., Khalouki, H., & Billah, M. E. (2025). A comprehensive evaluation of machine learning techniques for forecasting student academic success. Journal of Electronics, Electromedical Engineering, and Medical Informatics, 7(1), 1-2. Acharya, A., & Sinha, D. (2014). Early prediction of students performance using machine learning techniques. International Journal of Computer Applications, 107(1), 37-43. Al-Fairouz, E. I., & Al-Hagery, M. A. (2020). Students performance: From detection of failures and anomaly cases to the solutions-based mining algorithms. International Journal of Engineering Research and Technology, 13(10), 2895-2905. Alam, A., & Mohanty, A. (2022). Predicting students’ performance employing educational data mining techniques, machine learning, and learning analytics. In International Conference on Communication, Networks and Computing (pp. 166-177). Springer. Alruwais, N., & Zakariah, M. (2023). Student-engagement detection in classroom using machine learning algorithm. Electronics, 12(3), 731. Bulusu, S., Kailkhura, B., Li, B., Varshney, P. K., & Song, D. (2020). Anomalous example detection in deep learning: A survey. IEEE Access, 8, 132330-132347. Gao, Y. (2025). Deep learning-based strategies for evaluating and enhancing university teaching quality. Computers and Education: Artificial Intelligence, 7, 100362. Ghanim, J., & Awad, M. (2025). An unsupervised anomaly detection in electricity consumption using reinforcement learning and time series forest-based framework. Journal of Artificial Intelligence and Soft Computing Research, 15(1), 5-24. Huang, A. Y., Lu, O. H., Huang, J. C., Yin, C. J., & Yang, S. J. (2020). Predicting students’ academic performance by using educational big data and learning analytics: Evaluation of classification methods and learning logs. Interactive Learning Environments, 28(2), 206-230. Hussain, S., & Khan, M. Q. (2023). Student-performulator: Predicting students’ academic performance at secondary and intermediate level using machine learning. Annals of Data Science, 10(3), 637-655. Hussain, S., Gaftandzhieva, S., Maniruzzaman, M., Doneva, R., & Muhsin, Z. F. (2021). Regression analysis of student academic performance using deep learning. Education and Information Technologies, 26(1), 783-798. Issah, I., Appiah, O., Appiahene, P., & Inusah, F. (2023). A systematic review of the literature on machine learning application of determining the attributes influencing academic performance. Decision Analytics Journal, 5, 100204. Kaggle. (n.d.). Students Performance in Exams. Retrieved from https://www.kaggle.com/datasets/spscientist/students-performance-in-exams/data Kamalov, F., Sulieman, H., & Santandreu Calonge, D. (2021). Machine learning based approach to exam cheating detection. PLOS ONE, 16(8), e0254340. López-García, A., Blasco-Blasco, O., Liern-García, M., & Parada-Rico, S. E. (2023). Early detection of students’ failure using machine learning techniques. Operations Research Perspectives, 11, 100292. Nassif, A. B., Talib, M. A., Nasir, Q., & Dakalbab, F. M. (2021). Machine learning for anomaly detection: A systematic review. IEEE Access, 9, 78658-78700. Pallathadka, H., Wenda, A., Ramirez-Asís, E., Asís-López, M., Flores-Albornoz, J., & Phasinam, K. (2023). Classification and prediction of student performance data using various machine learning algorithms. Materials Today: Proceedings, 80, 3782-3785. Pek, R. Z., Özyer, S. T., Elhage, T., Özyer, T., & Alhajj, R. (2022). The role of machine learning in identifying students at-risk and minimizing failure. IEEE Access, 11, 1224-1243. Riestra-González, M., del Puerto Paule-Ruíz, M., & Ortin, F. (2021). Massive LMS log data analysis for the early prediction of course-agnostic student performance. Computers & Education, 163, 104108. Shitaya, A. M., Wahed, M. E., Ismail, A., Shams, M. Y., & Salama, A. A. (2025). Predicting student behavior using a neutrosophic deep learning model. Neutrosophic Sets and Systems, 76, 288-310. Vaidya, A., & Sharma, S. (2024). Anomaly detection in the course evaluation process: A learning analytics–based approach. Interactive Technology and Smart Education, 21(1), 168-187. Wang, G., Han, S., Ding, E., & Huang, D. (2021). Student-teacher feature pyramid matching for anomaly detection. arXiv preprint arXiv:2103.04257.

Recently, student’s academic performance prediction has become an increasingly prominent research topic in the field of Educational Data Mining (EDM). The prediction of student’s academic performance aims to explore information that is beneficial to the learning process of student. Therefore, accurate prediction of student’s academic performance provide benefits for education institutions to improve the quality of their institutions by improving the learning process of students. In predicting the student’s academic performance, the problem of high dimensional dataset is often faced in the datasets which significantly impacts the accuracy of student academic performance prediction. This paper proposed Genetic Algorithm based Feature Selection (GAFS) along with selected single classifier for classification in order to improve the accuracy in predicting student academic performance. Kaggle dataset is used in this paper and two phase of experiment have been conducted, single classifier without GAFS, and single classifier with GAFS. Results from the experiments show that, the accuracy of the proposed GAFS for classification makes an impressive performance in predicting student academic performance in terms of accuracy compare to existing techniques.

The academic performance of student is influenced by several factors. Studies have been conducted in the field of educational data mining to find out what all the factors are that have an effect on a student's academic performance. There are many factors that can have an effect on a student's academic success, but our study aims to find the main factors that can have an effect on a student's academic performance. Predicting student performance becomes more difficult due to the large volume of information in academic databases. The responsibility of the teacher increases, she must be attentive to the activities and behavior of the student. To facilitate the work of a teacher, this study attempted to identify some factors that affect the academic performance of students. Our study could bring benefits and impacts to students, educators/teachers, and tutoring institutions. The factors investigated in this study are: student base knowledge, socio-economic status, college/school environment, environment pollution, teachers’ support, parental/family support, friend circle and health.

Lung and colon cancers are among the most prevalent and lethal malignancies worldwide, underscoring the urgent need for advanced diagnostic methodologies. This study aims to develop a hybrid deep learning and machine learning framework for the classification of Colon Adenocarcinoma, Colon Benign Tissue, Lung Adenocarcinoma, Lung Benign Tissue, and Lung Squamous Cell Carcinoma from histopathological images. Current approaches primarily rely on the LC25000 dataset, which, due to image augmentation, lacks the generalizability required for real-time clinical applications. To address this, Contrast Limited Adaptive Histogram Equalization (CLAHE) was applied to enhance image quality, and 1000 new images from the National Cancer Institute GDC Data Portal were introduced into the Colon Adenocarcinoma, Lung Adenocarcinoma, and Lung Squamous Cell Carcinoma classes, replacing augmented images to increase dataset diversity. A hybrid feature extraction model combining MobileNetV2 and EfficientNetB3 was optimized using the Grey Wolf Optimizer (GWO), resulting in the Lung and Colon histopathological classification technique (MEGWO-LCCHC). Cross-validation and hyperparameter tuning with Optuna were performed on various machine learning models, including XGBoost, LightGBM, and CatBoost. The MEGWO-LCCHC technique achieved high classification accuracy, with the lightweight DNN model reaching 94.8%, LightGBM at 93.9%, XGBoost at 93.5%, and CatBoost at 93.3% on the test set. The findings suggest that our approach enhances classification performance and offers improved generalizability for real-world clinical applications. The proposed MEGWO-LCCHC framework shows promise as a robust tool in cancer diagnostics, advancing the application of AI in oncology.

Alzheimer’s disease is a progressive neurodegenerative disorder and the most common cause of dementia, leading to memory loss and cognitive decline. Early diagnosis is essential to improve patient outcomes, yet remains difficult due to subtle structural changes in the brain during the early stages of the disease. This study introduces a hybrid deep learning and machine learning framework for classifying different stages of Alzheimer’s disease using four-class OASIS-1 MRI scan data. The proposed method extracts discriminative features using the ConvNeXt-Base model from both cropped brain regions and segmented tissue images (gray matter, white matter, cerebrospinal fluid). These features are then classified using multiclass Support Vector Machines, applying One-vs-One and One-vs-Rest strategies. Our approach achieves high classification performance and demonstrates strong potential for improving early-stage detection and stage-specific diagnosis of Alzheimer’s disease.

Transient simulation is the key to ensure the safe and stable operation of power systems. After large disturbances, power systems may experience transient rotor angle instability and short-term voltage instability, which share similar electrical characteristics but require different control strategies. Therefore, distinguishing the dominant instability mode (DIM) is crucial. In this paper, a novel hybrid deep learning (DL) framework is proposed to achieve accurate DIM identification by fully extracting local-global features from electricity data. In the proposed framework, an advanced graph neural network (GNN) optimized by random sampling and aggregation is constructed. It can better capture local features than traditional GNN and help to improve model generalization ability. Further, important global features are mined by employing transformer network with self-attention mechanism to improve DIM identification accuracy. In addition, the vital discrete fault features are also embedded into the neural networks to improve the performance. The proposed method improves upon existing DL models by providing a solution for integrating multi-level features extracted by different schemes and addressing the limitation of focusing on a single instability mode. Case studies conducted on an 8-machine 36- bus system and Northeast China Power Grid verify the superiority of the proposed method over other state-of-the-art DL methods.

To address the challenges of false negatives and false positives of small objects and the difficulty of fine‐grained behavior recognition in complex traffic scenarios, this paper constructs a hybrid deep learning framework based on image detection to synergistically improve multi‐object localization accuracy and semantic understanding capabilities. The framework first uses a combination of Gaussian and bilateral filtering for denoising, enhancing input quality and improving detection sensitivity for small objects. In the detection phase, the YOLOv5s (You Only Look Once 5s) model is used as the baseline. The Convolutional Block Attention Module (CBAM) attention mechanism is applied to enhance the representation of key features. K‐means clustering is used to adaptively generate prior anchor boxes that match the scale distribution of objects in traffic scenarios. The CIoU (Complete Intersection over Union) loss function is also used to optimize bounding box regression accuracy, improving small object detection performance while maintaining model lightweight. To achieve fine‐grained semantic understanding, a two‐branch classification network is designed. The attribute branch uses the ConvNeXt‐Tiny (Convolutional Next‐Generation Tiny) structure to extract static appearance features, while the event branch utilizes the nonlocal operations module to capture dynamic contextual dependencies. Weighted fusion of these two features enables joint recognition of attributes and behaviors. A GNN‐CNN (Graph Neural Network‐Convolutional Neural Network) hybrid classification module is also constructed. The GNN models the spatiotemporal interactions between vehicles, while a lightweight CNN extracts local texture features. These features are adaptively fused using the Squeeze‐and‐Excitation (SE) attention mechanism, and a softmax classifier performs traffic behavior judgment. Experiments show that the YOLOv5s‐CBAM model achieves a mean average precision (mAP) of 0.55 for detecting extremely small objects (< 16 × 16). In the overloaded vehicle detection task, the GNN‐CNN module achieves accuracy and recall of 0.92 and 0.90, respectively. This hybrid deep learning framework provides reliable technical support for automated traffic inspections. It improves the accuracy and stability of small object detection and fine‐grained event recognition in complex traffic scenarios. Its modular design and strong scalability make it widely applicable and conducive to promoting intelligent transportation towards higher levels of automation.

The tremendous development and rapid evolution in computing advancements has urged a lot of organizations to expand their data as well as computational needs. Such type of services offers security concepts like confidentiality, integrity, and availability. Thus, a highly secured domain is the fundamental need of cloud environments. In addition, security breaches are also growing equally in the cloud because of the sophisticated services of the cloud, which cannot be mitigated efficiently through firewall rules and packet filtering methods. In order to mitigate the malicious attacks and to detect the malicious behavior with high detection accuracy, an effective strategy named Multiverse Fractional Calculus (MFC) based hybrid deep learning approach is proposed. Here, two network classifiers namely Hierarchical Attention Network (HAN) and Random Multimodel Deep Learning (RMDL) are employed to detect the presence of malicious behavior. The network classifier is trained by exploiting proposed MFC, which is an integration of multi-verse optimizer and fractional calculus. The proposed MFC-based hybrid deep learning approach has attained superior results with utmost testing sensitivity, accuracy, and specificity of 0.949, 0.939, and 0.947.

This study examines the impact of online gaming on the academic performance of Pokhara university business students in Kathmandu Valley. Academic performance of business students is the dependent variable. The selected independent variables are online games, social media network, time spent on social media, mobile phones, friends and people connection and internet addictions. The primary source of data is used to assess the opinions of the respondents regarding the impact of online gaming on academic performance of Pokhara university business students in Kathmandu Valley. The study is based on primary data with 123 respondents. To achieve the purpose of the study, structured questionnaire is prepared. The correlation coefficients and regression models are estimated to test the significance and importance of online gaming on academic performance of Pokhara university business students in Kathmandu Valley. The study showed that online gaming has a positive impact on academic performance of business student indicating that playing informative online gaming leads to increase in the level of academic performance of business students. Likewise, the result also revealed that social media network has a positive impact on academic performance of business students. It means that use social media network helps in improving student’s academic performance. Moreover, time spent on social media has a positive impact on academic performance of business students indicating that more the time students spent on social media for the purpose, better would be the academic performance of the students. Moreover, mobile phones has a positive impact on academic performance of business students. It indicates that use of mobile phones helps in improving academic performance of business students. Further, the result shows that friends and people connection has a positive impact on academic performance of business students which means that connecting with friends and people through social media helps students to improve their academic performance.

This study examines the relationship between commuting time and students' academic performance, further explores the impact of commuting time on academic performance and its underlying mechanisms, and analyzes strategies to mitigate the negative effects of commuting time on student academic performance. The data source for this study is the 2014-2015 China Education Panel Survey (CEPS). This study first employs linear regression to analyze the impact of commuting time on students' academic performance (Chinese, mathematics, and English scores). It then utilizes the PROCESS procedure to examine the mediating effects of psychological fatigue and learning engagement between commuting time and academic performance. Finally, an interaction term between commuting time and commuting mode is constructed, and hierarchical regression is applied to test the moderating effect of commuting mode on the relationship between commuting time and students' academic performance. (1) According to linear regression findings, each additional minute of commuting time correlates with a 0.04-point decrease in Chinese scores, a 0.032-point decline in math scores, and a 0.05-point drop in English scores. (2) The negative impact of commuting time on students' academic performance exhibits significant variation based on individual student and family factors. (3) Learning engagement and psychological fatigue partially mediate the relationship between commuting time and student performance. (4) Commuting mode significantly moderates the impact of commuting time on students' academic performance. Commuting time significantly impedes students' academic performance. Learning engagement and psychological fatigue partially mediate the relationship between commuting time and students' academic performance in this population. Commuting mode significantly moderates the effect of commuting time on students' academic performance. Compared to passive commuting modes, active commuting intensifies the negative impact of commuting time on academic performance. To mitigate the negative effects of commuting time on students' academic performance, school administrators should optimize campus locations, implement proximity-based enrollment policies, enhance boarding services, and provide psychological counseling and academic support to students.

The study analysed the relationship between parenting styles and academic performance of senior high school students in the Kumasi Metropolis. Convenience and simple random sampling techniques were used to select the schools and 376 respondents, respectively. Pearson’s Correlation Coefficient Matrix and multiple ordinary least square were used to estimate the impact of parenting styles on academic performance. The study found that authoritative parenting style had greatest significant positive impact on academic performance of the students (β = 0.104, SE = 0.011, t = 9.539, p < 0.001), followed by authoritarian style (β = 0.044, SE = 0.011, t = 3.971, p < 0.001). However, permissive style had no significant impact on students’ academic performance (β = 0.042, SE = 0.025, t = − 1.682, p = 0.093). Neglecting style had significant negative impact on students’ academic performance (β = 0.072, SE = 0.011, t = − 6.740, p < 0.001). Following from the study findings, it is recommended that the Schools in collaboration with the Parent Teacher Association should organise guidance and counselling programs for parents to sensitise them on various parenting styles and their ramifications on academic performance of students. Also, Class Counsellors’ professional capabilities to manage the psycho-social problems of students are imperative for an improved academic performance of the latter.

The purpose of the study was to identify the effect of the face-to-face tutoring program of undergraduate students in Administration on academic performance, as well as its level of student satisfaction with the program. The research questions were: (a) what are the aspects of the institutional tutoring program that impact on the academic performance of the students? (B) What is the satisfaction stage from the tutored about the tutoring institutional program attended? and (c) what effect does the institutional mentoring program have on the academic performance of administration students during their scholar journey? The methodology used in this study was based on the quantitative approach, non-experimental with a transactional design. 63 eight semester students participated, and they attended the tutoring program. The Pearson correlation test was used for the aspects included in the tutoring program and the academic performance of the students. The results indicate that there is no significant relation between the aspects that the tutorial program covers and the academic performance, however, is a high rate student satisfaction with the tutoring program. In the results of Pearson´s correlation test between the aspects of the program and the general average, no significant relation was found in the aspects covered by the program. As a conclusion, the institutional tutoring program does not have a statistically significant relation in the students' academic performance, in contrast to student satisfaction, which was favorable to the tutoring program.&#x0D; Key words: Effect, student, academic performance, satisfaction, tutor&#x0D; Code JEL: I20, I21, I23&#x0D; Received: 24/06/2019. Accepted: 10/02/2020. Published: 27/04/2020

Introduction. The relevance of the study is determined by the active involvement of young students in the use of digital technologies and the lack of a holistic view of the cognitive characteristics, indicators of educational activity and academic performance of students with different parameters of digital competence and Internet use. The purpose of the article is to study cognitive and stylistic characteristics, indicators of educational activity and academic performance of schoolchildren and students with different parameters of digital competence and Internet use. Materials and methods. The study involved students from schools and universities in St. Petersburg, Moscow, Novosibirsk, and Yekaterinburg (total sample: N=583). Methods: a survey using a questionnaire to collect socio demographic information (also contained a question regarding the ways of using the Internet and a question to determine self-assessment of academic performance), psychodiagnostic method. Methods: "Digital Competence Index" (G.U. Soldatova, T.A. Nestik, E.I. Rasskazova, E.Yu. Zotova); "Thinking Style" (A.K. Belousova); "Schoolchild Academic Activity Questionnaire" (A.A. Volochkov); "Student Academic Activity Questionnaire" (A.A. Volochkov). Statistical processing of the results was carried out in the Python programming language using the Google Colaboratory development environment, using exploratory factor analysis, Student's t-test. KEYWORDS Results. The structure of relationships between digital competence and ways of using the Internet with cognitive characteristics and academic performance of students has been established. It has been determined that the expression of digital competence in schoolchildren and students in the field of content, software, consumption is combined with more effective assimilation of educational material and is reflected in the ability to critically evaluate information, apply it in practice, coordinate joint activities. Compared with schoolchildren, students have higher rates of academic performance (p≤0.01) and learning ability (p≤0.01), associated with their cognitive strategies of organizational orientation (p≤0.01) and active practices of using the Internet for educational purposes (p≤0.01). digital competence, thinking styles, educational activity, academic performance Conclusion. The study contributes to the understanding of the profile of cognitive-style characteristics, indicators of educational activity and academic performance of schoolchildren and students with different parameters of digital competence and use of the Internet.